AI DIY Episode 2 Social Introduction Bot

12/14/21 • 66 min

Ken Cheng is a comedian, professional poker player and presenter of Game Over, Human, a BBC Radio 4 program about game players and the AIs that beat them. His choice of AI is a robot that will handle social networking situations for him, introducing him to people at parties and telling anecdotes to make him look good. This is an idea that the event industry is really excited about, with AI Matchmakers that can advise all the delegates about who they want to be speaking to. If that's too corporate for you, how about a dating coach, or Microsoft's TrueMatch software to pick good teams of online gameplayers.

Hosted on Acast. See acast.com/privacy for more information.

Hosted on Acast. See acast.com/privacy for more information.

Previous Episode

AI DIY Episode 1 Robot Politician

Professor Age Chapman is an expert on the data pipeline that companies use to develop AI and Data Science products. Her choice of future AI is a robot politician - but what even is a politician for? It's unsurprising that science fiction is full of robot leaders and politicians — see Isaac Asimov's story "Franchise" about how democracy works in a world controlled by an all-powerful computer. However, it may surprise you to know that in real life, most people approve of replacing their politicians by robots and that an AI chatbot ran against Vladimir Putin in the Russian general election in 2018.

Hosted on Acast. See acast.com/privacy for more information.

Next Episode

AI DIY Episode 3 Roboaudit

Dr Pamela Ugwidike is an expert on the use of AI in justice systems and is particularly interested in how AI and data bias impact criminal justice. She wants to make an AI that can audit other AIs for bias to make sure that all Artificial Intelligence always operates in the interest of humanity. There's currently a huge amount of interest in AI ethics and biases in the research community, and we are discovering many ways in which our

data and our computations can lead to unfairness.

Perhaps the most famous case of AI bias comes from the field of criminal justice. The COMPAS algorithm (used in US courts to predict whether a defendant would go on to reoffend) was twice as likely to incorrectly label black people as reoffenders than white people. While there are lots of advice and programming toolkits aimed at helping human developers to eradicate bias from the AI products that they are building, IBM's Watson OpenScale is the only AI platform that claims to detect and correct biases in its own operation. Although is that just IBM's marketing department being a bit biased?

Hosted on Acast. See acast.com/privacy for more information.