Amazon expects to be the next leader in generative artificial intelligence (AI), according to CEO Andy Jassy.
Speaking to shareholders at the firm’s virtual AGM, held on Wednesday, May 24, Jassy told investors the company is at the very early stages of using generative AI and expects it to offer ‘high potential’ to both customers and the business. Over the last nine months there had been a new inflection point for generative AI within Amazon, Jassy remarked. ‘It’s still very early days, but it’s moving quickly’ he said.
The $1.24 tn company will be incorporating AI more substantially in three layers. The bottom layer of AI looks at training models to run inference or predictions on computers. The middle layer focuses on offering large foundational models onto which people can build generative AI. The top layer looks at the applications that will be built on top of the foundational models.
Jassy said: ‘While we pursue a number of those applications, many more will be built by external companies and developers. We’re quite optimistic that a lot of those applications will be built on top of Amazon Web Services. So it’s very early days in generative AI, it’s very high potential: we’re investing quite a bit and expect to be a leader.’
Misuses of AI
Despite the headway being made in AI, shareholders remained concerned over the risks surrounding advanced technologies. During the AGM, shareholders were able to vote on item 23, which requested that Amazon publish a report on customer use of certain technologies, including using AI to assist with purchases and sales.
One of the biggest concerns raised by shareholders was around Amazon’s ability to market and sell facial-recognition software to the US government, according to the AGM report.
‘Human and civil rights organizations are concerned facial-surveillance technology may violate civil rights by unfairly and disproportionately targeting and surveilling people of color, immigrants and civil society organizations,’ said John Harrington, a shareholder who raised the proposal. He called on the board to release a report on the use of technologies no later than September 1, 2023.
Responsible use of AI
The board recommended voting against item 23 as it ‘fails to acknowledge or address’ the measures taken to ‘enhance our technology and relies on dated claims and mischaracterizations’.
The board said: ‘While we understand the concerns over potential misuse, we believe these are effectively addressed through the policies and procedures we have adopted and that we continue to advance with input from internal and third-party partners and stakeholders.
‘Amazon is committed to the responsible use of our AI and machine-learning products and services. We have been consistent and proactive in our efforts to address concerns and mitigate the risk of misuse through policy and advocacy efforts.’
The results of the AGM found 4.5 mn shareholders voted against the proposal while only 2.7 mn voted for the release of a report.