In recent years, the UK has seen more use of AI in government, from deciding benefits to enforcing laws.
Artificial Intelligence systems have benefits, but an investigation raised concerns about their impact on citizens’ lives and potential biases.
Challenges and Concerns in AI Use
The investigation revealed that various Whitehall departments are employing Artificial Intelligence and algorithms in a “chaotic and frequently unregulated approach.”
A Member of Parliament claimed that the Department for Work and Pensions made an error by utilizing an algorithm. This mistake resulted in numerous individuals losing their benefits.
It demonstrates the significant impact that AI decisions can have on real-life situations. This shows how Artificial intelligence decisions can have severe consequences in real life.
Bias in AI Systems
Bias in AI systems is another pressing issue. The police used a face recognition tool. This tool had more mistakes in identifying Black faces than White faces, which shows a problem with the technology’s use.
AI Implementation Across Departments
Many government departments, including eight Whitehall departments and some police forces, use Artificial Intelligence. For instance, the Home Office has utilized an algorithm to detect potential “sham marriages.”
Still, it has disproportionately flagged individuals from specific countries, such as Albania, Greece, Romania, and Bulgaria, raising concerns about potential discrimination.
The Impact of Training Datasets
The root of these problems often lies in the training datasets used to teach Artificial Intelligence systems. Biased information in datasets leads to biased AI systems.
This highlights the importance of having good data and ethical AI development. Developers, especially those working on big language models like ChatGPT, need to understand how these models work entirely.
AI’s Potential in the NHS
The UK’s Prime Minister, Rishi Sunak, champions the potential benefits of AI in the National Health Service (NHS). It also has the potential to accelerate diagnosis, improve the accuracy of tests, and optimize bed management. However, experts caution that public services should carefully implement AI as it is not a one-size-fits-all solution.
AI’s Unintended Consequences
AI’s unintended consequences are a significant concern. In Holland, they used artificial intelligence to detect fraud in childcare benefits. However, incorrect judgments wrongly pushed many families into poverty. Mr. Sunak is holding a global summit at Bletchley Park to discuss the dangers of AI to people.
The CEO’s Perspective
Shameem Ahmad, CEO of the Public Law Project, sees the benefits of AI for society but also warns about its downsides. Urgent action is required to prevent the routine and possibly unlawful application of opaque automated systems in life-altering ways, where individuals may lack the means to seek redress when these processes go awry.
AI in Government and Policing
The government has been using Artificial Intelligence in the civil service and police. Examples of this include electronic passport gates and automatic number plate recognition.
Lately, people have been advocating for the government to use AI more openly and consistently.
Recently, there has been a push for increased transparency and uniformity in the government’s use of AI. This involves ensuring that individuals affected by these tools comprehend their usage and can question them when necessary.
Push for Transparency
The Cabinet Office is urging government departments and law enforcement agencies to disclose the potential of AI. They want to share this information publicly through an “algorithmic transparency reporting standard.”
Currently, six entities have disclosed information about their initiatives. A recent request for information revealed that eight government departments are utilizing Artificial intelligence technology. This has caused concerns regarding transparency and accountability.