White House AI safety pact ABC hears differing views

Sep 27

ABC’s Rachel Scott reports Trump and technology executives signing a voluntary, nonbinding AI accord. Trump favors industry self-policing; Dario Amodei says the risks are real and the response still under discussion. The report replays Bill Gates’s earlier warning and asks Trump about calling concerns a hoax; he says he means Democrats’ conduct. Amodei declines a further question. The speakers’ conflicting assessments are preserved, and the agreement is not presented as a binding safeguard.

Gates urges lawmakers to regulate AI and warns that sufficiently powerful systems could drive events causing a billion deaths. This is his warning about possible consequences, not a forecast that such deaths will occur or an announcement of legislation. Kan's longer September 28 package explores the danger of malicious biological and cyber uses and calls for government safeguards and monitoring. It places Gates's warning alongside other expert concerns and Trump's emphasis on the industry's economic scale and U.S. leadership. The closing preserves Gates's support for beneficial AI applications in education, medicine and climate. Catastrophic harm is a warned-of possibility, not a quantified forecast or an assertion that a specific attack is planned. AUF1's September 29 report replays Gates's warning about catastrophic harm and the combination of malicious intent with powerful AI tools. It then quotes editor Stefan Magnet alleging coordinated fearmongering and speculating that such warnings could push smaller developers out in favor of large companies. The report presents this as his interpretation and supplies no evidence of coordination. Gates's scenario remains a risk warning; the longer Kan report's benefits and safeguards discussion is preserved alongside this contrasting account. AUF1's September 30 account replays the same interview warning and excerpts Magnet's September 29 evening response at greater length. Magnet argues that fear creates demands for security, enabling regulations affordable only to large corporations and controlled information from centralized providers. This is his attributed prediction about the warnings' consequences, without proof of coordination or of an enacted regulatory or market outcome.