AI reinforces racism

Unite delegate Susan Matthews highlights AI discrimination threat

Reading time: 5 min

Unite delegate Susan Matthews gave a barnstorming speech as she moved a composite on the impact of AI in the workplace on Tuesday (September 10) at TUC Congress.

Susan acknowledged that the scale of the impact of AI across all sectors was “far-reaching and only beginning to be understood”.

She highlighted the work Unite has done to identify jobs at risk and ensure the upskilling of workers to prevent redundancies through job displacement.

“There are real challenges on this, if the technology is used in a way that increases social isolation, and reduces face to face interaction, rather than helps to connect people,” she noted.  

“But I want to focus on is a major threat of AI that gets little or no mention at all,” Susan went on to say. “It is about how the entrenched, systemic discrimination of our society is being recreated, replicated and reinforced through AI.”

Susan noted how new technologies are enabling and amplifying racism, sexism, homophobia, ableism and a number of other forms of discrimination.  

“But it is not new — the form may be different, but the outcome is the same,” she added. “It’s the same old discrimination and oppression. It is becoming clearer and clearer this is happening across different industries and workplaces, and society.”

Susan emphasised that it was vital that trade unions give training to workplace reps to raise awareness of this particular impact of AI.

Susan went on to recount how police in the US have arrested the wrong Black person because AI “had been ‘trained’ that Black people looked the same”.

“Here, in the UK software is being used to award exam marks that penalise students from Black and Asian minority backgrounds and low-income neighbourhoods,” she noted. “We see software that targets and reinforces prejudices, delivering socially divisive and hate-filled online media content to people.”

Susan pointed to research finding that software models are replicating racist stereotypes. For example, language models have found that software carries out linguistic discrimination against those who spoke African-American English, who were assigned worse jobs and are more likely to be branded ‘lazy’.

“And we know that technology is being increasingly used to control workers through observation, with the low paid and outsourced Black and Asian Ethnic Minority workers being the most at risk to surveillance work,” she added.

“There is sometimes the false narrative that technology is above human prejudices and therefore better. That is not true,” Susan went on to say.

“Congress, after years of fighting and fighting against discrimination, there is now the real threat of it being further embedded against women, Black and Asian ethnic minority, disabled and LGBT+ workers before we are properly alive to what discrimination is going to look like.”

While Susan acknowledged that “technology can and does include the opportunities for human advancement and for a better life and society for all”, she added, “This is not a technology problem. It is at root a human, social, economic, deprivation, industrial and political problem.”

“And it needs human, social, economic, industrial and political solutions,” she continued.

“Congress, we have many challenges to our diversity and inclusivity at the moment, and many threats that AI brings,” Susan warned. “We need to make sure we meet this one as trade unions in terms of training our reps to ensure that AI doesn’t discriminate in the workplace.”

Concluding with a quote by civil rights leader Malcom X, Susan said, “Education is a passport to the future – for tomorrow belongs to those who prepare for it today.”

To thundering applause, she added, “AI is our future – so let us prepare.”

By Hajera Blagg

Photo by Mark Thomas