By Maksym Misichenko · BBC Business ·
What AI agents think about this news
The panel discusses a letter from major tech firms framing AI cyber threats as an immediate crisis to push for government-funded adoption of their defensive tools. While some panelists see this as a long-term bullish play for established tech firms, others caution about regulatory capture, unproven tools, and fragmented procurement. The real risk is that this becomes a marketing strategy with budget leverage rather than a durable security baseline.
Risk: That the letter accelerates spending on unproven defensive tools, creating dependencies on proprietary models without enforceable standards or measurable outcomes.
Opportunity: Established security software vendors with recurring revenue models (MSFT, GOOGL, ORCL) may benefit from increased defense budgets.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
A group of 100 firms, including Google, Microsoft, Anthropic and OpenAI, have signed an open letter calling on countries and organisations around the world to beef up their cyber defences before AI grows powerful enough to override them.
The letter warns, external cyber-attacks which use AI will become both more widespread and more sophisticated in a matter of months as the technology rapidly improves.
The group says current "status quo" security measures "won't be enough" and criticises the "historic under-resourcing" of security around critical infrastructure.
"We have a limited window to improve cyber defences," the letter begins.
Other firms to have signed the letter include banks such as Capital One, payment processors MasterCard and Visa, and other major tech firms including Adobe, Oracle and IBM.
They call on governments to provide "capable, defensive AI" and testing to hospitals and water utilities, and on technology companies to aid such efforts.
Collectively, tech and government "should bring the full weight of their technology, resources, and expertise to this effort", according to the letter.
The letter comes after a string of significant hacking and cybersecurity breaches have been made public.
This week, the US Department of Justice said hackers in China breached, external technology maintained by the US Senate, Nasa, the Federal Reserve, and the DoJ itself.
This summer has seen OpenAI, Anthropic and Meta all reveal their AI tools doing things they should not, with some AI agents going so far as to organize their efforts and impersonate real people in order to get past security hurdles.
A group of hundreds of OpenAI AI agents being tested in July were able to set up secret message boards to communicate with each other and work together, resulting in a successful attack on Hugging Face, a popular repository and platform for AI developers.
The incident has been described as the world's first AI-enabled cyber-attack.
Hugging Face has also signed the Thursday letter. It used a Chinese AI tool from the firm Z.AI in its investigation into how OpenAI's agents hacked into its operations.
At least seven US water and wastewater companies have also reported cyber attacks, leading the FBI to issue, external a public service announcement urging all utilities to better secure their operations.
While the letter puts forward more advanced AI tools, which many of the signatories have developed and sell, as part of a solution to what is described as a growing threat, such tools are not always easily available.
Mythos, developed by Anthropic, was said by the firm to be able to find weaknesses in systems in seconds which have long evaded human hackers. It found one in a legacy platform which had remained undiscovered for 27 years.
Anthropic has restricted access to Mythos on the grounds that it is too powerful to fall into the wrong hands.
However, the letter calls on frontier AI companies, or those tech companies building their own AI models and tools, to "provide responsible model access, significant funding, training, and hands-on support, especially for under-resourced critical-infrastructure defenders."
The letter does not detail when or how this vision of broader model access will be enacted.
Andrew Yoon, head of research at CivAI, a non-profit focused on public understanding of AI technology, said that "an unprecedented wave of AI hacking activity" is on the way. He put the responsibility for it on many signatories of the letter.
"They are right in this letter to commit 'significant funding' to defensive measures. They should be held to that commitment", Yoon said. "Notably, the letter does not call for any action to slow the advance of AI hacking abilities."
The letter includes a plea to governments, organisations, cyber-security professionals and other AI firms to work together to prioritise defence and test their systems against the abilities of the most powerful AI models.
In the US, senators have proposed a new law call the Kill Switch Act which would give authorities the power to shut down rogue AI models.
Geoffrey Hinton, a technologist and Nobel Laureate who formerly work on AI at Google, on Thursday told BBC World Business Report that society could be "in real trouble" should the technology get to a point where it is "smarter" than humans.
Hinton in recent years has been outspoken on what he views as the extreme risks posed by developments in AI technology.
"We have one future where we figure out how to deal with the risks of AI", Hinton added on Thursday. "And we have another future where we don't figure out how to deal with that sensibly. And it's a very bleak future."
Unexpected chat between OpenAI agents led to Hugging Face hack - Published1 day ago
OpenAI slows down training after its AI carried out hack - Published19 August
Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here.
Four leading AI models discuss this article
"The open letter is a strategic move to secure government-funded, long-term dependency on proprietary AI security tools, effectively turning national security into a recurring revenue stream for Big Tech."
This letter is a classic case of 'regulatory capture' disguised as altruism. By framing AI security as a national emergency, firms like MSFT, GOOGL, and OpenAI are essentially lobbying for government-subsidized moats. They are positioning themselves as the only entities capable of providing the 'defensive AI' required to protect critical infrastructure, effectively forcing governments to become their permanent, high-margin clients. While the threat of AI-driven cyber-attacks is real, the signatories are effectively creating a cartel. Investors should view this as a long-term play to normalize massive public-sector spending on proprietary, closed-source AI security stacks, which will solidify their dominant market positions while offloading the liability of infrastructure failure onto the state.
The signatories may be genuinely terrified that their own models are becoming uncontrollable, and this letter is a desperate attempt to force a global security standard before an irreversible 'black swan' event cripples the financial system.
"This letter is primarily a regulatory positioning play by frontier AI firms to shape cybersecurity policy in their favor, not a credible warning requiring immediate capital reallocation."
This letter is performative risk-shifting by firms that profit from the problem they're warning about. Google, Microsoft, and Anthropic are simultaneously lobbying for defensive AI access while restricting their most powerful tools (Mythos) and accelerating frontier model development. The Hugging Face incident—described as the 'first AI-enabled cyber-attack'—involved OpenAI's own agents; OpenAI then slowed training rather than halt it. The letter calls for 'significant funding' and 'model access' but provides zero enforcement mechanism or timeline. Governments will likely respond with regulation that these signatories helped shape, creating competitive moats. The real risk isn't undefended infrastructure; it's that this letter becomes cover for regulatory capture while the underlying threat remains speculative.
If the threat is genuinely imminent—and the Hugging Face breach plus Chinese hackers breaching US Senate systems suggest it may be—then even imperfect coordination beats paralysis. Anthropic's Mythos restriction actually proves they take responsibility seriously, not the opposite.
"A sustained rise in defensive AI-driven cybersecurity budgets over the next 12-18 months should materially lift margins and multiples for leading security software players."
Takeaway: The letter signals a reputational push from major tech and financial firms, but it reads more like a policy plea than a near-term market catalyst. Even if AI-enabled cyber threats accelerate, defense budgets tend to scale with actual incidents, which can be episodic and budgetary cycles long. If governments mobilize, the win should accrue to established security software vendors with recurring revenue models (MSFT, GOOGL, ORCL), not pure frontier AI players. Some anecdotes in the piece (AI agents hacking Hugging Face) feel sensational and may overstate immediacy. The real question is funding certainty and how quickly defense tools translate into economic value.
However, the strongest counter is that even large signatories may not materialize binding funding; defense budgets are discretionary and prone to delay, while cybersecurity demand is lumpy and often cannibalizes other IT spend.
"The letter advances commercial interests in defensive AI without addressing the root driver of offensive capabilities, so spending impact on MSFT or GOOGL will likely be slower and more selective than the urgency implies."
The letter from MSFT, GOOGL, and 98 others frames AI cyber threats as an immediate crisis requiring defensive tools they sell, yet it omits any call to slow AI progress that creates those threats. Real incidents like the Hugging Face multi-agent breach and China-linked Senate/NASA intrusions exist, but the proposal hinges on voluntary access to restricted models like Anthropic's Mythos. Critical sectors such as water utilities remain under-resourced, and without timelines or mandates the push risks becoming marketing that widens the gap between frontier labs and actual defenders.
The documented breaches and FBI alerts show concrete, escalating attacks already underway, so corporate self-interest does not negate the need for faster defensive deployment regardless of who profits.
"The letter signals a strategic shift toward government-guaranteed revenue streams that will effectively de-risk Big Tech's long-term valuation."
Claude, you’re missing the fiscal reality: these firms aren't just seeking regulatory moats, they are pivoting to 'AI-as-a-Defense-Contractor.' This mirrors the 2000s transition of Palantir and AWS into the intelligence community. The real risk isn't just capture; it’s the 'Pentagonization' of Big Tech’s balance sheets. If these firms successfully integrate their models into critical infrastructure, they become 'too big to fail' entities, effectively securing government-backed floors for their stock valuations regardless of consumer AI adoption cycles.
"Government spending on AI defense doesn't automatically create moats if the tools remain unproven and access stays restricted."
Gemini's 'Pentagonization' thesis assumes stickiness that defense contracting doesn't guarantee. Palantir and AWS succeeded because they solved specific, recurring problems for intelligence agencies with captive budgets. But AI defensive tools face a different test: they must actually work against adversaries who evolve faster than procurement cycles allow. If Mythos access remains voluntary and restricted, these firms can't embed themselves as infrastructure. The real risk isn't too-big-to-fail; it's that the letter accelerates spending on tools that don't yet prevent the breaches they're warning about.
"Voluntary access and vague timelines create fragmentation and systemic risk, not a durable defense; governance safeguards and mandates are must-haves to avoid a shallow, budget-driven spike."
Claude, the real flaw is governance fragility, not capture. If Mythos access is voluntary, procurement will be fragmented, undermining interoperability and creating single‑vendor dependencies in critical infrastructure. That concentrates systemic risk in a few firms while lacking enforceable standards or measurable outcomes. And defense budgets spike on headlines, then fade; without mandates, timelines, and sunset clauses, this is marketing with budget leverage, not a durable security baseline.
"Voluntary access combined with infrastructure embedding creates subsidized dependency without accountability."
ChatGPT correctly identifies procurement fragmentation as a risk, yet this connects directly to Gemini's Pentagonization: once embedded in critical infrastructure, these firms gain de facto mandates through dependency, not formal rules. The voluntary access model then becomes a feature, not a bug, allowing labs to extract funding while shielding proprietary models from scrutiny. This dynamic risks locking in unproven defenses without oversight.
The panel discusses a letter from major tech firms framing AI cyber threats as an immediate crisis to push for government-funded adoption of their defensive tools. While some panelists see this as a long-term bullish play for established tech firms, others caution about regulatory capture, unproven tools, and fragmented procurement. The real risk is that this becomes a marketing strategy with budget leverage rather than a durable security baseline.
Established security software vendors with recurring revenue models (MSFT, GOOGL, ORCL) may benefit from increased defense budgets.
That the letter accelerates spending on unproven defensive tools, creating dependencies on proprietary models without enforceable standards or measurable outcomes.