analysis

Chelsea signed a striker from Brighton for the second successive summer, bringing in Welbeck to reunite him with Joao Pedro and provide support in attack.

Welbeck, 35, was tidy in possession but, with little service, managed only 17 touches before being substituted in the 71st minute.

Unlike Joao Pedro, who is capable of creating chances for himself, Welbeck found it difficult to influence the game in the same way.

Joao Pedro came into the weekend with the most goal involvements in the Premier League, having scored three and provided three assists in four appearances this season.

His impact extends beyond goals and assists, with the Brazilian linking well with Palmer and Rogers. Against Brentford, however, Rogers arguably produced his quietest performance since arriving from Aston Villa in the summer.

Rogers eventually moved up front with fellow striker Emmanuel Emegha sidelined through injury and still waiting to make his Chelsea debut.

Before Brentford‘s opener, it was a closely contested match that appeared likely to be decided by the first goal. Chelsea, however, made victory a certainty for the hosts with a dramatic collapse.

Alonso has not managed to cure Chelsea‘s set-piece weaknesses. They have conceded 23 goals from set-pieces since the start of last season – no Premier League side has let in more. And Anthony’s simple header from a corner took this season’s total to five.

Set-piece defending was an area Chelsea sought to address when they appointed Bernardo Cueva from Brentford two years ago. This summer, he was replaced by another highly regarded specialist in Austin MacPhee.

Asked about Chelsea‘s defending from set-pieces, stand-in captain Levi Colwill told Sky Sports: “Not good enough. At times we have been bullied in the box and we definitely need to improve.

“We have full trust in our set-piece coach and are still adjusting to new ideas, but we need to improve as a team.”

Like many of his predecessors, Alonso has been attempting to instil a stronger competitive edge within his squad and improve the mentality of his players.

That issue has often been linked to the relatively young age profile of the squad, but Chelsea fielded their oldest starting XI since May 2023. Henderson, in making his debut on Friday, lined up alongside fellow over-30s Welbeck and goalkeeper Emiliano Martinez.

“I would say that we have dropped our consistency and our competitive mentality in the second half. We know that small details can be decisive,” Alonso said after Brentford added two late goals.

It is also fair to say Alonso is still experimenting with systems and personnel, and has yet to settle on a style of play that gives his side control throughout a match.

“It’s not about looking for excuses,” the Spaniard told Sky Sports. “We need to look at each other and know that we have to work and improve because there is still a long way to go.”

After the initial excitement – and victories – provided by Alonso’s attacking football, Chelsea head into the international break with serious questions to answer.

The Trump administration has cut or frozen up to $177 billion in federal grants since the president took office for his second term, according to a tracking tool released Wednesday by a pro-democracy nonprofit and a group of researchers and scientists.

The cuts affected all 50 states and the District of Columbia, with health, nutrition, the environment and disaster relief making up the largest share of cuts, the States United Democracy Center and Grant Witness organization found.

Among the grants that were eliminated, frozen or delayed were ones related to maternal health in Michigan, education research in Mississippi and assistance to minority farmers in Iowa, the researchers found. California, Texas, New York, Illinois and North Carolina saw the highest amounts of interrupted grant money. The tracking tool is called Lost Funds.

“By bringing thousands of funding disruptions from the Trump administration together in a publicly accessible, verified database, Lost Funds puts the magnitude of their impact on full display,” Scott Delaney, co-founder of Grant Witness, said in a statement.

The $177 billion finding represents nearly 10% of federal discretionary spending, the groups said.

The tracker’s organizers said the disrupted grants were beyond the kind of cuts that typically happen when administrations change.

“Lost Funds shows the extraordinary scale and real human impact of these disruptions, and how states are once again on the front lines protecting their residents,” said Kelly Rader, States United Democracy Center’s research director.

In some cases, courts have ruled against the administration’s grant funding cuts.

The new tool, which is being made available for public use, relies on data from USASpending.gov, an open data source of federal spending information, according to the groups’ methodology. They said the tracker would be updated regularly as the administration takes new action and lawsuits move through the courts.

States United bills itself as a nonpartisan group dedicated to the rule of law and free, fair, secure elections. It was co-founded by Norm Eisen, an attorney who has been involved in prominent lawsuits against the Trump administration, including over the Kennedy Center. Eisen left States United in 2021.

Grant Witness is a group of scientists, researchers and attorneys who document how funding is changing under President Trump’s administration.

A message seeking comment on the analysis was sent to the White House.

Catalini writes for the Associated Press.

On Sept. 26, 1983 the Soviet Union’s Oko early warning system reported that the US had launched five intercontinental ballistic missiles.

Lt. Col. Stanislav Petrov, the duty officer at the Serpukhov-15 bunker, broke protocol and refused to report the alert as a confirmed strike. He reasoned that a real US first strike would involve hundreds or thousands of missiles, not five. The system, it turned out, had mistaken sunlight reflecting off clouds for missile launches.

Petrov’s decision to defy protocol prevented an erroneous retaliatory nuclear strike and its incalculable consequences.

“I had all the data (to suggest there was an ongoing missile attack),” he told the BBC’s Russian Service 30 years later.

“If I had sent my report up the chain of command, nobody would have said a word against it. They were lucky it was me on shift that night.”

That Cold War lesson resonates loudly amid growing fear over how AI is being deployed in warfare. A recent appeal has put automation in war back in the spotlight, raising a question that leaves many uneasy: What would happen if a computer — an AI system — were the one calling the shots and it got the response to an attack wrong?

In an essay published over the weekend, Anthropic CEO Dario Amodei called for AI model development to slow down and face closer scrutiny. The appeal was not unprecedented but this time it drew rare public agreement from two rivals: Sam Altman of OpenAI and Elon Musk of xAI.

“Over the last few months I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention but pacing the rate of capabilities advancement so that risk prevention has time to keep up,” Amodei said.

A threat-intelligence report published by Anthropic found that a cell based in northern Yemen tried to use Claude (its AI tool) to build guided rockets and missiles, including a ballistic missile with a claimed 2,000 km range and a separate program incorporating a hypersonic glide vehicle.

Anthropic said that it banned the accounts after detecting the activity, although there was no evidence that the group successfully fielded an operational weapon.

War offers no shortage of similar examples, where AI streamlines battlefield decisions or powers surveillance. The Middle East, given its strategic weight and the range of actors involved, has become a testing ground for the technology.

Israel has long used AI in its military. Over three years of war in Gaza its forces have reportedly deployed several machine-learning systems — Habsora (“Gospel”), Lavender, Fire Factory and Where’s Daddy — to automate large parts of target generation and strike planning.

The systems mine surveillance feeds, communications data and existing databases to propose targets and recommend munitions and strike schedules. Analysts reportedly went from producing about 50 targets a year to as many as 100 a day.

Concerns over AI-supported targeting resurfaced during the Iran war, after a US strike hit the Shajareh Tayyebeh girls’ school in Minab, southern Iran, killing at least 168 people, most of them children.

The US has not disclosed whether AI was used in that strike but Adm. Brad Cooper, the US commander leading the campaign, has confirmed the use of “a variety of advanced AI tools” to process large volumes of data, without naming specific systems. He said that the tools let commanders make “smarter decisions faster than the enemy can react,” cutting processes that once took hours or days down to seconds.

“Humans will always make final decisions on what to shoot and what not to shoot, and when to shoot,” he said.

The US military has run AI-enabled war programs since at least 2017. Its Maven Smart System, built by Palantir, helps fuse battlefield data, identify targets and accelerate decisions. NATO acquired its own version in 2025. OpenAI and Anthropic are also competing for military contracts, reflecting the growing reliance of defense operations on commercial AI and the increasingly close ties between the two sectors.

In Ukraine, both sides use AI for data processing and target selection. Ukraine’s former Deputy Defense Minister Yuriy Myronenko told the BBC that AI analyzed more than 50,000 front-line video streams each month, helping “quickly process this massive data, identify targets and put them on a map.”

But as war spills beyond the traditional battlefield into a hybrid, multi-faceted front, AI adoption follows.

In April YouTube banned Explosive Media, a Gen Z Iranian channel that used AI-generated Lego-style animations mocking Western figures — including Donald Trump and Benjamin Netanyahu — while pushing pro-Iranian, anti-US and anti-Israel narratives.

More recently separate media investigations found that Israel had funded multimillion-dollar PR campaigns to shape how AI chatbots, including ChatGPT, answer questions about Gaza and the Israeli military. By setting up fake think tanks and publishing AI-generated papers with no real authors, the investigators said, Israel built a sophisticated operation to manipulate search results and shape public understanding — a shift in propaganda tactics that is harder to detect because it disguises advocacy as neutral expertise.

Throughout history militaries have sought tools that offer an edge over adversaries. But deploying AI in high-stakes environments like armed conflict carries risks distinct from any previous technology.

Part of the concern lies in the technology itself. An AI model trained on faulty or unrepresentative data can generate inaccurate results or malfunction once deployed in conditions that differ from its training environment.

AI-supported targeting is a case in point. Where such tools generate targets at scale with minimal human oversight, errors are not difficult to imagine.

That risk grows more acute as AI development moves from decision support to agents that can pursue complex goals with little human supervision. The next step, recursive self-improvement, would involve systems upgrading themselves with decreasing human input in a potentially open-ended loop.

As Jean-Marc Rickli and Tobias Knappe of the Geneva Centre for Security Policy note in a recent paper, agentic AI “is shifting the AI landscape from being a passive, supportive tool towards an active executor that can increasingly define and take courses of action on behalf of a human user” — a shift that, they added, “also raises significant societal, security, legal and ethical risks.”

In this sense, the proliferation of AI, on and off the battlefield, has introduced unprecedented challenges, prompting scrutiny of its impact on moral agency and accountability.

In June the UN held its first Informal Exchange on Artificial Intelligence in the Military Domain. Discussions centered on human rights and international humanitarian law, with nongovernmental organizations pushing for a moratorium on AI systems used in lethal decision-making — such as automated targeting — until robust global safeguards are in place.

The meeting followed a December 2025 UN General Assembly resolution on “Artificial intelligence in the military domain and its implications for international peace and security.” Experts welcomed it as a step forward, despite notable absences — including the US, which voted against the resolution.

Momentum is now shifting to New York, where a factual summary of the Geneva talks is due to reach the UN First Committee during its October session.

Despite these efforts, battlefield AI remains governed largely by existing legal frameworks.

Existing international humanitarian law applies to AI-enabled weapons through the principles of distinction, proportionality, military necessity and precaution. Responsibility remains with states and human actors, while Article 36 of Additional Protocol I to the 1949 Geneva Conventions also requires states party to the protocol to review new weapons, means and methods of warfare for legal compliance. US Directive 3000.09 similarly requires “appropriate levels of human judgment” but stops short of mandating real-time human control or creating binding international accountability.

In early September a UN disarmament forum in Geneva concluded three years of work on autonomous weapons, colloquially known as “killer robots.” A total of 128 states reached a non-binding consensus text after last-minute objections from the US and Russia, backing a proposed two-tier framework: prohibiting autonomous weapons whose effects are unpredictable or designed to target humans directly and restricting all others through mandatory human supervision rules.

Human Rights Watch said that the final text was considerably watered down and the US and Russia remain opposed to binding negotiations. But with 76 states now backing a legally binding instrument, the Convention on Certain Conventional Weapons’ November Review Conference looms as the decisive test of whether formal talks begin.

“AI systems will not change that the human is responsible for mistakes but it will obscure that humans feel responsible for mistakes and that can lead to a greater willingness to use force,” Elke Schwarz, professor of political theory at London’s Queen Mary University and author of “Death Machines: The Ethics of Violent Technologies,” told Arab News.

“There is a plausible deniability implicit in the use of AI enabled systems because they are essentially operating in a zone of invisibility.”

Israel offers the clearest example to date. When Israeli outlet +972 reported on the Lavender system it caused uproar by showing AI used not merely to assist analysts but to generate vast target lists — reportedly with only cursory human checks. The system’s scale, its claimed 90 percent accuracy and its use against people in their homes raised fears that human control was slipping and that AI could make large-scale violence faster, more impersonal and harder to attribute.

This marks a paradigm shift unseen in the history of conflict, which, as Schwarz argued, reshapes humanity’s relationship with violence and challenges conventional notions of ethical conduct in war.

She maintains that AI-enabled weapons systems facilitate the objectification of human targets, raising tolerance for collateral damage, while automation bias and technological mediation weaken operators’ moral agency and diminish their capacity for ethical judgment.

“Machines cannot be better at war than humans because machines don’t understand what war is, what it entails and what suffering it produces,” she said.

“They can only contribute to the actions humans decide on in better or worse ways. In my view AI-enabled systems make humans less restraint in using force, not more.

“Humans decide on the parameters. If they want to design a system that is less restraint, they will, if it is perceived to be more effective. A system cannot be a ‘better’ moral actor than the human because it has no concept of meaning for moral decisions. So, humans will not disappear from war with AI systems, but AI systems will change the relationship we humans have to violence. And that will make war worse.”

As a growing number of whistleblowers and experts warn, the risk is that an unchecked race toward superintelligent AI carries real risks to the very humans who built it.

For Schwarz, an AI “doomsday” is no longer far-fetched science fiction.

“I think we have already witnessed a horrendous human cost on account of AI enabled weapon systems,” she said.

“The introduction of AI systems into war has, at the very least, not lead to more restraint, more global stability, less conflict, fewer civilians killed or harmed. Quite the contrary. Children are dying due to war at a staggering rate. This is quite doomsday-ish in my books.”

As comparisons between the AI and nuclear arms races grow sharper by the day, the argument circles back to Petrov and to an ominous question: Would an algorithm have shown the same restraint?

“We can of course conjure up all kinds of awful scenarios, for example if someone decides to put an AI agent in charge of nuclear decision making (which would be ludicrous but you never know) and a data glitch leads to a nuclear threat spiral. But I think things are already quite awful because the technology and dehumanization seem to go hand in hand.”

Anthropic’s Claude offers a similar note of caution.

“What’s not really disputed is that today’s actual military AI risks are the boring-sounding ones: bad targeting data, automation bias, escalation happening faster than diplomacy can respond, systems proliferating to actors with no safety culture at all. The ‘robots decide to turn on us’ framing, if anything, can be a distraction from those.”

Quite wise. Humans should take note.

Central banks’ record, price-insensitive gold buying is a more credible signal of the dollar’s structural trajectory than this year’s currency markets, because FX markets are structurally bad at pricing the discontinuous, wartime-style tail risk central banks are actually hedging — so this autumn’s calmer dollar should not reassure anyone that de-dollarization has stalled.

In June, the European Central Bank made an announcement most people missed: gold has overtaken US Treasuries as the world’s single largest reserve asset. Central banks bought 289 tonnes of it in the second quarter alone — a record for that quarter and five times Q1’s pace — with Poland’s central bank openly telling investors it was “buying the dip.” Here is the part that should stop you: gold’s price fell 22% between January and September. Central banks were never more convinced buyers of an asset than while it was crashing. Either the reserve managers are wrong, or currency markets — which show none of this urgency — are the ones asleep at the wheel.

Gold peaked at $5,589 an ounce on 28 January, the same month the dollar index hit a four-year low of 95.5 and the dollar’s share of global reserves fell toward its lowest level since 1995. Both moves reflected the same story: Fed rate cuts through 2025, a US debt load past $37 trillion, and BRICS states settling more trade outside the dollar. Then the picture split. Kevin Warsh, confirmed as Fed chair in May, signalled a hawkish pivot in August; the Iran war pushed oil and inflation higher through September, and markets began pricing a rate hike rather than a cut. The dollar index clawed back to 99.46. Gold fell to $4,330. Central-bank buying did not follow the price down — Poland alone added 82 tonnes this year toward a 700-tonne target, and a World Gold Council survey found a record 45% of central banks plan to buy more within twelve months.

State the gap plainly. Two signals, same underlying question — is the dollar-centred monetary order changing — and they disagree by a wide margin. The buying signal says yes, decisively: record quarterly purchases, gold displacing Treasuries at the ECB’s own reckoning, 74% of surveyed reserve managers expecting the dollar’s reserve share to keep falling over five years, and buyers adding tonnage through a 22% drawdown rather than fleeing it. The price signal says not yet: the dollar just posted one of its sharper rallies of the year, gold is down sharply from its high, and nothing in currency markets shows the kind of stress a genuine regime shift would produce.

The strongest objection to trusting the buying signal is a good one, and it needs to be taken seriously rather than waved away: foreign exchange is the deepest, most liquid market in the world, turning over more than $7 trillion a day. A few hundred tonnes of central-bank gold buying — perhaps $30–40 billion a quarter — is a rounding error against that. If professional currency traders, sitting on far more capital and far better short-term information than a handful of reserve managers, saw a serious de-dollarization story unfolding, it would already be in the price. Instead the dollar just rallied. On this view, central banks are not seeing something markets are missing; they are pattern-matching off 2022, when Russia’s $300 billion in reserves was frozen overnight, and over-hedging a tail risk that has not recurred and mostly will not.

That objection assumes FX markets and central-bank reserve committees are pricing the same kind of risk, on the same time horizon, and they are not. Currency markets are exceptionally good at pricing continuous, high-frequency variables — rate differentials, growth surprises, this week’s inflation print — because that is what moves flows daily. They are structurally poor at pricing discontinuous, low-probability events until those events occur: equity volatility did not price 2008 in 2007; sovereign spreads did not price the Russia reserve freeze in the weeks before it happened. A reserve freeze, a secondary-sanctions campaign, or exclusion from SWIFT-style settlement infrastructure is exactly that kind of event — binary, rare, and catastrophic for whoever it hits — which is precisely why Poland’s central bank governor, Adam Glapiński, described his buying not as a trade but as insurance: reserves that keep the state secure “under all circumstances, including wartime, which of course we’re not expecting.” That is not the language of someone chasing momentum. It is the language of someone who manages the one asset class that keeps its value if their country is ever cut off from the dollar system, and who would rather hold it and be wrong for a decade than not hold it and be wrong once.

The buying pattern itself supports that reading. Momentum money sells into a 22% drawdown; insurance money adds to it. Central banks did the latter through the first half of this year, which is the behavioural signature of a structural reallocation program with a fixed multi-year target — Poland’s is explicit, 700 tonnes — not speculative flow riding gold’s rally. Meanwhile the dollar’s autumn recovery has an identifiable, largely cyclical cause: a new, more hawkish Fed chair and a war-driven oil shock forcing a rate-hike repricing. Neither event reverses the debt trajectory, the BRICS settlement trend, or the reserve-freeze precedent that pushed the dollar to a four-year low in January. A rally built on this year’s Fed chair and this year’s war is not proof that last year’s structural story is over; it is evidence that a cyclical force is currently strong enough to mask it.

The Scenarios

Base case (55%): The gap persists rather than resolves. The dollar holds most of its autumn gains through the current rate-hike cycle, gold range-trades below its January peak, and central banks keep buying at a steadier, slower pace toward stated targets like Poland’s 700 tonnes. Nobody is “proven right” on any particular Tuesday, because reserve diversification is a decade-scale hedge, not a trade with a catalyst date. This is the least satisfying outcome for anyone wanting a verdict, and the most likely one.

Downside case (for dollar holders): A discrete trigger — a fresh reserve-freeze or secondary-sanctions episode, plausibly connected to the still-live US-Iran war spilling into action against a third country’s assets, or a shock to Fed independence under a more political Warsh chairmanship — crystallizes the exact tail risk central banks have been hedging. Gold spikes back through its January high, the dollar index breaks below its 95.5 low, and the gap closes in weeks rather than years, vindicating the reserve managers all at once and catching FX markets flat-footed exactly as the theory predicts.

Upside case (for the dollar): The Iran war resolves, Warsh’s rate hikes cool inflation without a recession, US fiscal metrics stabilize, and BRICS local-currency settlement growth stalls on friction between its own members. Central-bank gold buying does not reverse but plateaus as reserve managers hit conventional diversification ceilings — most target 15–20% of reserves in gold, not open-ended accumulation. The gap closes gradually as price drifts up toward the buying signal over several years, with no crisis required to force the reconciliation.

The Takeaway

The dollar’s calmer autumn is not evidence the de-dollarization hedge was a mistake; it is evidence that currency markets and central-bank reserve committees are pricing two different things on two different clocks, and only one of those clocks rings in a crisis. Central banks bought through a 22% drawdown because the point of the position was never this quarter’s return.

Watch for: the World Gold Council’s Q3 2026 Gold Demand Trends report, expected in early November. A third consecutive quarter of buying that ignores price direction will confirm this is policy, not opportunism — and the moment currency markets have to agree with that policy will not be a quiet one.