Artificial intelligence is a new technology with the potential to solve some of the world’s most pressing problems, but it also carries frightening risks if it grows out of control and harms humanity. 

This mix of risk and opportunity applies to the press, as it does to most modern institutions. Researchers at Mila, led by scientific director Hugo Larochelle, are focused on maximizing the opportunities.

In September, Mila named Hugo Larochelle, former head of Google’s AI lab, as its new scientific director. “I take on these responsibilities with deep enthusiasm and a sense of duty,” he said.

At Mila, that duty includes safety measures to limit AI’s short- and long-term risks.

Larochelle will provide the keynote address at WPFC’s November symposium. 

In a release on news his Mila appointment, he said the work carried out at Mila paves the way for scientific breakthroughs while making a positive contribution to society.

Dario Amodei, CEO of AI research company Anthropic, recently said there is a 25 per cent chance the future of AI “will go really, really badly.”  

Amodei had been asked about the “doom number” — tech shorthand for AI’s potential to destroy humanity or trigger a catastrophe. He also said there’s a 75 percent chance that “things go really, really well.”

Newsrooms can use AI tools to analyze vast amounts of data, for example. 

But others are using the technology to manipulate public opinion, create deepfakes, and spread disinformation. This weakens trust in the media and creates safety issues for journalists, increasing threats and attacks on the work they do. 

Mila works on how best to use AI and machine learning to improve systems and sectors while simultaneously trying to address the negative challenges of AI, including disinformation and deepfakes. 

Founded by AI leader Yoshua Bengio, Mila is the world’s largest academic research centre for deep learning. Its goal is to make machine-learning algorithms more trustworthy, safe, and fair — principles that directly benefit the media and a free press. If news and public affairs consumers trust the information they consume, they are more likely to support press freedom.

Mila partnered with UNESCO to publish Missing Links in AI Governance, a call for AI regulation. Regulation, they argue, would help citizens with their AI interactions and for the media, it would require transparency, including labelling AI-generated content to prevent deception.