Billionaire Bill Gates warned that Artificial Intelligence is already «powerful enough» that, if deployed maliciously, it could even trigger events capable of causing one billion deaths.
Speaking on the program «Meet the Press», Bill Gates argued that self-regulation by technology corporations themselves is insufficient and that state intervention, statutory law, and substantive oversight are required.
His statements come during a period in which Washington and the AI industry are attempting to determine what the next chapter should look like for a sector expanding at unprecedented velocity, while warnings regarding potentially extreme consequences multiply concurrently.
«There has never been a stronger weapon»
NBC News released an excerpt Friday from the interview of Bill Gates on «Meet the Press», during which the billionaire and philanthropist was asked to comment on warnings by industry executives that AI could, under extreme conditions, threaten humanity itself.
«AI is certainly powerful enough to lead to events that could cause a billion deaths», stated Bill Gates.
As he noted, «there has never been a weapon as powerful as the combination of people with malevolent intent using the most advanced AI tools».
The warning fits into an increasingly intense debate over whether the most sophisticated Artificial Intelligence systems can turn hazardous when deployed by governments, criminal organizations, or rogue actors.
The intervention by Dario Amodei and the debate on tapping the «brakes» on AI
Bill Gates was also questioned regarding a recent essay by the CEO of Anthropic, Dario Amodei, who called for decelerating the development of the most advanced AI frontier models.
Dario Amodei argued that top artificial intelligence laboratories should moderate their pace of progress and «cooperate voluntarily to establish standards» surrounding AI.
Bill Gates was emphatic that this does not suffice.
«Nobody thinks self-regulation is enough», he remarked, commenting on the prospect of AI companies themselves assuming the primary role in mitigating hazards.
According to him, legislation regulating Artificial Intelligence is «absolutely» required.
«You need law enforcement and politicians inside the discussion on how guardrails and monitoring mechanisms should be structured», he noted.
AI moves faster than governments
Dario Amodei has likewise advocated for government intervention in the AI arena.
However, he highlighted the primary structural dilemma: governments advance far more slowly than technology.
«Unfortunately, passing laws can take time, while AI evolves very rapidly», he observed.
The question remains what concrete measures the federal government of the US will enact.
Thus far, the posture in Washington is characterized by sharp dissension, with Republican leaders in Congress appearing hesitant regarding the necessity for stringent oversight.
Donald Trump has likewise positioned himself firmly against strict restrictions on AI, dismissing appeals for a tighter regulatory framework as a «scam».
What comes next
His remarks take on added weight since, according to multiple news accounts, a meeting has been scheduled for Tuesday between President Donald Trump, Speaker of the House Mike Johnson, and leading executives from the Artificial Intelligence sector.
Thus far, however, the precise agenda of the meeting has not been clarified.
Mike Johnson: «Lock them in a room and solve it»
Mike Johnson previously indicated that he would like Donald Trump and AI CEOs to «get into a big room, close the door, and figure out a solution».
The preceding week, he also stated to journalists that there should be «deliberate and serious discussion regarding corporate responsibility for maintaining safety and what role, if any, should be assigned to the government».
It remains undisclosed which specific heads of AI enterprises will attend the gathering.
The meeting takes place mere days after prominent figures of the tech industry, including Elon Musk and the head of OpenAI Sam Altman, attended an official dinner at the White House on Thursday.
Will Congress ultimately legislate on AI?
Low legislative speed in Washington despite warnings
Mike Johnson and Senate Majority Leader John Thune have exhibited little urgency so far regarding the passage of legislation regulating AI.
Mike Johnson downplayed fears that the technology could lead to the extinction of humanity, characterizing them as a «scam».
John Thune, by contrast, has favored some format of AI regulation for years and introduced a relevant legislative bill in 2023.
Lately, however, he appears less inclined to place the issue at the summit of the legislative docket.
In mid-September, John Thune told reporters that whether the Senate can pass AI legislation before the end of the year remains «fairly open».
«Accomplishing anything in the short term will be difficult, given all the other issues we have to address», he said.
Renewed push for bipartisan support
Despite reservations, John Thune appears to continue evaluating the issue.
Speaking to Axios on Wednesday, he mentioned that Senator Amy Klobuchar, who co-sponsored his previous legislative initiative on AI, is working to secure broader backing from Democrats before the bill is reintroduced in Congress.
Under the version tabled in 2023, the legislation would mandate, among other provisions:
certification of critical AI systems,
heightened transparency standards for content synthesized through generative AI,
statutory obligations for companies to execute risk assessments for Artificial Intelligence systems,
expanded federal research dedicated to detecting AI-generated content,
studies examining safety guardrails that must be deployed across the sector.
It remains uncertain, however, whether this specific measure or any alternative bill can marshal adequate votes in Congress and, even more critically, avert a potential veto by Donald Trump.
John Thune nevertheless stated to Axios that he believes Donald Trump might prove more receptive to certain safety guardrails for AI than he projects publicly.
The broader debate on AI safety intensifies
Resignation of Jacob Coxon unleashed a fresh wave of reactions
The debate surrounding Artificial Intelligence safety and regulation has escalated noticeably over recent weeks.
On September 9, former Anthropic researcher Jacob Coxon publicly announced his resignation from the firm.
Concurrently, he delivered an exceptionally alarming statement:
«The people building AI genuinely believe it could kill us all by the end of the decade».
His statement unleashed a fresh wave of scrutiny and anxiety regarding development boundaries for advanced AI frontier models.
Evan Hubinger: Greater than 10% probability of a catastrophic scenario
The warning was reinforced by Evan Hubinger, alignment science lead at Anthropic.
Evan Hubinger stated that Jacob Coxon is «correct» and that AI executives «genuinely believe AI could kill all humans».
He stated that he considers there to be a probability «above 10%» that AI will cause the extinction of humanity «within the next decade».
Simultaneously, he warned that Anthropic «does not yet possess a plan for solving the alignment problem in the event of superintelligence and is not clearly on a trajectory leading to a solution».
These declarations amplified public outcry and intensified pressure for the enactment of an oversight framework.
Google, Anthropic, and OpenAI explore a united front on safety
Within this climate, Google, Anthropic, and OpenAI have begun coordinating moves and formulating plans to establish a working group centered on AI safety.
Concurrently, Sam Altman and Dario Amodei placed heavy emphasis this week on the imperative for international cooperation and cross-border standards for Artificial Intelligence during presentations before the UN.
Dario Amodei argued that decisions regarding the technology «must be shaped through democratic processes and by governments accountable to the people they serve».
Deloitte: Three-tier model for operational deployment of AI
In sponsored content, Deloitte presents a three-stage framework for integrating Artificial Intelligence into operational workflows across health insurance organizations.
The methodology rests upon three axes: optimization, redesign, and reinvention, aimed at extracting substantial added value from AI.
www.bankingnews.gr
Readers’ Comments