TCS+ | The Up&Up Group on the hidden cost of AI
Loading player...
Companies large and small are pouring capital into AI projects, chasing the promise of efficiency, speed and scale. But as Jason Harrison, chief operating officer of The Up&Up Group, argues in this episode of TechCentral’s TCS+, the upfront price tag tells only a fraction of the story – and many South African boards are signing cheques without fully understanding what they’re buying.
Harrison uses The Up&Up Group’s own experience in experimenting with and implementing AI to glean insight into the gap between the large promises by the technology (and its Silicon Valley pundits) and harsh realities of stalling projects in enterprises, especially at the integration layer.
The conversation digs into the costs that rarely make it into a chief financial officer’s spreadsheet:
• Policy and governance, Harrison argues, are not soft considerations – they are line items, often substantial ones.
• The costs of error: the reputational damage, financial exposure and legal risk when copyrighted material slips into outputs or autonomous agents go off-script.
• Agents that run amok, with companies then only discovering the problem once the cloud bill lands.
Beyond the balance sheet, Harrison flags AI’s energy footprint as a societal cost the industry is still reckoning with.
AI adoption is accelerating despite these risks and Harrison describes the current moment as a kind of nuclear arms race, driven by share-price pressure and a deep fear of being last. Much of today’s AI spend, he suggests, is Fomo (“fear of missing out”) dressed up as strategy. For tech leaders trying to make sober decisions inside a hype cycle, separating signal from noise has become a leadership skill in its own right, he says.
Suggested solutions include a “test and learn” philosophy where many small, inexpensive AI experiments are run throughout an organisation before viable instances are scaled appropriately.
Harrison cautions against one-size-fits-all deployments and argues that governance must sit close to the work rather than only in the boardroom. Quarter-to-quarter measurable objectives matter, he says, but only if they live inside a long-term strategy.
Despite his sober-minded view on AI’s high costs, Harrison still has an optimistic perspective on the technology and its potential to transform society, especially on the African continent. While the developed world is using AI to get answers, Harrison suggests African organisations may end up using it to learn how to think – a subtle but important distinction, and a timely note for any leader weighing their next AI investment.
Don't miss this discussion.
Harrison uses The Up&Up Group’s own experience in experimenting with and implementing AI to glean insight into the gap between the large promises by the technology (and its Silicon Valley pundits) and harsh realities of stalling projects in enterprises, especially at the integration layer.
The conversation digs into the costs that rarely make it into a chief financial officer’s spreadsheet:
• Policy and governance, Harrison argues, are not soft considerations – they are line items, often substantial ones.
• The costs of error: the reputational damage, financial exposure and legal risk when copyrighted material slips into outputs or autonomous agents go off-script.
• Agents that run amok, with companies then only discovering the problem once the cloud bill lands.
Beyond the balance sheet, Harrison flags AI’s energy footprint as a societal cost the industry is still reckoning with.
AI adoption is accelerating despite these risks and Harrison describes the current moment as a kind of nuclear arms race, driven by share-price pressure and a deep fear of being last. Much of today’s AI spend, he suggests, is Fomo (“fear of missing out”) dressed up as strategy. For tech leaders trying to make sober decisions inside a hype cycle, separating signal from noise has become a leadership skill in its own right, he says.
Suggested solutions include a “test and learn” philosophy where many small, inexpensive AI experiments are run throughout an organisation before viable instances are scaled appropriately.
Harrison cautions against one-size-fits-all deployments and argues that governance must sit close to the work rather than only in the boardroom. Quarter-to-quarter measurable objectives matter, he says, but only if they live inside a long-term strategy.
Despite his sober-minded view on AI’s high costs, Harrison still has an optimistic perspective on the technology and its potential to transform society, especially on the African continent. While the developed world is using AI to get answers, Harrison suggests African organisations may end up using it to learn how to think – a subtle but important distinction, and a timely note for any leader weighing their next AI investment.
Don't miss this discussion.
