ZDNET56%

Today's challenge: Working around AI's fuzzy returns and questionable accuracy 56%

By Joe McKendrick49%

8/5/2024, 8:00:15 AM

BS Summary: This article contains 25 faulty reasoning types, including Hasty Generalization, Optimism Bias, and Pessimism Bias, with Negativity Bias as the most egregious example at 20.9% saturation with 136 hits. Analysis detected 1,252 faulty-reasoning hits from 652 analyzed words, generating a BS Score of 53.3% and a BS Rank of 56% (9,501 of 21,194 articles). This article is worse (more manipulative) than 55.20% of the article peer group.

It has become difficult to set realistic expectations about artificial intelligence -- and this could ultimately confuse efforts to understand the actual value of AI efforts. 
As the use of technology increases, it means changes in the career landscape for technology professionals, favoring more creative thinkers. 
That's the word from Ajay Malik, former head of architecture and engineering of Google's Worldwide Corporate Network, and currently CEO of Secomind.ai, who sees a rocky road ahead in the AI space. 
Perhaps one of the most challenging aspects of AI at this point is setting realistic expectations, he said in a recent podcast hosted by Thomas Erl, president of Arcitura Education. 
Also: Photoshop vs. 
Midjourney vs. 
DALL-E 3: Only one AI image generator passed my 5 tests 
For starters, there isn't enough measurement or awareness of the potential gains AI is delivering, Malik said. 
Decision-makers "want to be sure that all the information that they will use internally, or for interacting with customers, is accurate," he said. 
"How will companies measure the accuracy of what AI is doing? 
So AI did something, how do you always know it's accurate? 
How can you trust it 100%?" 
This weighs on how well business goals can be achieved through AI, Erl said. 
"If organizations are not successful or if they stumble, or if they invest in AI systems that end up resulting in loss instead of growth, that may postpone or change the outcome of how AI might impact their workforce. 
They might think, 'this didn't work out, let's go back to human workers.'" 
But the opportunity is real and we should prepare ourselves for whatever the impact will be. 
Unfortunately, there are no clear-cut "before-and-after" pictures that graphically illustrate the impact or accuracy of AI, Malik said. 
To address this, "they need to design built-in verification, built-in explainability, and built-in checks and balances to see if the AI's answer is correct." 
This includes "an alternative path, mechanism, model that provides a technique so that they can verify the answer. 
The key is understanding what exactly the AI system is producing, Malik advised. 
"Don't use AI as a black box that you depend upon without even thinking. 
We are not there today." 
In addition, businesses cannot rely on services such as ChatGPT, as responses need to be accurate and free of hallucinations. 
Instead, he advises, AI systems should have "checks and balances built in, verifying the answers, verifying the data, and offering explainability. 
There is a term for it called XAI, or explainable AI." 
There are also profound implications for technology-oriented career growth, Malik continued. 
"There is a big resource shift coming," he said. 
Those employees who use AI will become lot more valuable than the employees who do not use AI." 
AI's impact will be felt in the types of jobs and roles that will flourish in the months and years to come. 
"Even in software, even in programming, even in testing, a lot of those jobs will get eliminated -- not today, but over time," Malik predicted. 
"This is work which the AI can do -- very junior level work or very repetitive redundant level of work." 
This will especially apply to coder-level jobs, versus higher-level software engineering jobs, he continued. 
"Coders are just coding based on some known facts, and programming uses more thinking. 
In my own company, we see 20 to 25 times higher productivity because of using AI for supporting coding, for supporting meetings, meeting minutes, action items they can do a lot more with less people now." 
At the same time, there will be a shift toward "the thinkers, the problem solvers, the people who are creative," Malik added. 
"AI will take care of the labor, repetitive, or well-defined. 
But the creative humans will use AI to produce in high velocity and high quality and something really creative. 
That shift is coming." 
Article reasoning-pattern comparisonThis article: 0.0%Joe McKendrick: 0.6%ZDNET: 2.8%Confirmation Bias0.0%This article: 0.0%Joe McKendrick: 1.9%ZDNET: 1.6%Anchoring Bias0.0%This article: 1.7%Joe McKendrick: 3.1%ZDNET: 3.1%Availability Heuristic1.7%This article: 2.1%Joe McKendrick: 0.5%ZDNET: 1.0%Representativeness Heuristic2.1%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.5%Hindsight Bias0.0%This article: 5.5%Joe McKendrick: 4.1%ZDNET: 3.2%Overconfidence Bias5.5%This article: 0.0%Joe McKendrick: 3.4%ZDNET: 4.0%Framing Effect0.0%This article: 3.7%Joe McKendrick: 2.1%ZDNET: 1.4%Loss Aversion3.7%This article: 0.0%Joe McKendrick: 1.3%ZDNET: 0.6%Status Quo Bias0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.2%Sunk Cost Effect0.0%This article: 16.3%Joe McKendrick: 3.5%ZDNET: 5.0%Optimism Bias16.3%This article: 15.3%Joe McKendrick: 6.1%ZDNET: 1.3%Pessimism Bias15.3%This article: 20.9%Joe McKendrick: 10.5%ZDNET: 4.7%Negativity Bias20.9%This article: 8.3%Joe McKendrick: 2.2%ZDNET: 1.6%Self-Serving Bias8.3%This article: 0.0%Joe McKendrick: 0.9%ZDNET: 0.3%Fundamental Attribution Error0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Actor-Observer Bias0.0%This article: 0.0%Joe McKendrick: 0.2%ZDNET: 0.5%In-Group Bias0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.2%Out-Group Homogeneity Bias0.0%This article: 3.4%Joe McKendrick: 0.5%ZDNET: 3.9%Halo Effect3.4%This article: 0.0%Joe McKendrick: 0.2%ZDNET: 0.2%Horn Effect0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Dunning-Kruger Effect0.0%This article: 5.8%Joe McKendrick: 0.5%ZDNET: 1.5%Recency Bias5.8%This article: 0.0%Joe McKendrick: 0.3%ZDNET: 0.4%Primacy Effect0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Blind-Spot Bias0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Ad Hominem0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Straw Man0.0%This article: 8.3%Joe McKendrick: 10.9%ZDNET: 4.5%Appeal to Authority8.3%This article: 6.1%Joe McKendrick: 3.7%ZDNET: 1.5%False Dilemma6.1%This article: 15.2%Joe McKendrick: 3.2%ZDNET: 0.6%Slippery Slope15.2%This article: 0.0%Joe McKendrick: 0.3%ZDNET: 0.2%Circular Reasoning0.0%This article: 20.6%Joe McKendrick: 13.0%ZDNET: 6.5%Hasty Generalization20.6%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.5%Red Herring0.0%This article: 0.0%Joe McKendrick: 0.1%ZDNET: 0.5%Bandwagon0.0%This article: 9.4%Joe McKendrick: 1.7%ZDNET: 2.0%Appeal to Emotion9.4%This article: 2.0%Joe McKendrick: 0.9%ZDNET: 0.7%Begging the Question2.0%This article: 7.7%Joe McKendrick: 2.8%ZDNET: 1.5%Post Hoc (False Cause)7.7%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Tu Quoque0.0%This article: 4.4%Joe McKendrick: 0.4%ZDNET: 0.2%Burden of Proof4.4%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Appeal to Nature0.0%This article: 0.0%Joe McKendrick: 0.7%ZDNET: 0.2%Composition/Division0.0%This article: 10.1%Joe McKendrick: 5.1%ZDNET: 4.8%Anecdotal10.1%This article: 0.0%Joe McKendrick: 0.3%ZDNET: 0.1%No True Scotsman0.0%This article: 0.8%Joe McKendrick: 0.9%ZDNET: 2.4%Ambiguity (Equivocation)0.8%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Gambler’s Fallacy0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.2%Middle Ground0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Personal Incredulity0.0%This article: 0.0%Joe McKendrick: 0.3%ZDNET: 0.2%Special Pleading0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.1%Genetic Fallacy0.0%This article: 5.5%Joe McKendrick: 1.5%ZDNET: 1.0%Unattributed Quote5.5%This article: 0.5%Joe McKendrick: 0.2%ZDNET: 0.5%Quote-first Misdirection0.5%This article: 4.0%Joe McKendrick: 2.9%ZDNET: 6.3%Biased Writer Voice4.0%This article: 11.5%Joe McKendrick: 5.2%ZDNET: 4.0%Indoctrination11.5%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Politically Left Leaning Bias0.0%This article: 0.0%Joe McKendrick: 0.0%ZDNET: 0.0%Politically Right Leaning Bias0.0%This article: 3.1%Joe McKendrick: 1.4%ZDNET: 7.2%Attempt to Sell a Product or S…3.1%

652 words analyzed.

Speakers

2speakers78%attributed speech142writer words
Voice mapSelect a segment to jump to its words
Writer's voice • 10 words • 0.0% coverageWriter's voice • 26 words • 100.0% coverageWriter's voice • 20 words • 0.0% coverageWriter's voice • 32 words • 0.0% coverageAjay Malik • 30 words • 0.0% coverageWriter's voice • 3 words • 100.0% coverageWriter's voice • 2 words • 0.0% coverageWriter's voice • 11 words • 0.0% coverageAjay Malik • 17 words • 0.0% coverageAjay Malik • 23 words • 100.0% coverageAjay Malik • 11 words • 0.0% coverageAjay Malik • 11 words • 0.0% coverageAjay Malik • 6 words • 0.0% coverageThomas Erl • 14 words • 0.0% coverageThomas Erl • 39 words • 0.0% coverageThomas Erl • 13 words • 100.0% coverageWriter's voice • 16 words • 100.0% coverageAjay Malik • 18 words • 0.0% coverageAjay Malik • 24 words • 100.0% coverageAjay Malik • 18 words • 0.0% coverageAjay Malik • 13 words • 0.0% coverageAjay Malik • 14 words • 100.0% coverageAjay Malik • 5 words • 0.0% coverageAjay Malik • 20 words • 100.0% coverageAjay Malik • 21 words • 100.0% coverageAjay Malik • 11 words • 0.0% coverageAjay Malik • 11 words • 0.0% coverageAjay Malik • 9 words • 0.0% coverageAjay Malik • 18 words • 0.0% coverageWriter's voice • 22 words • 0.0% coverageAjay Malik • 25 words • 0.0% coverageAjay Malik • 20 words • 0.0% coverageAjay Malik • 14 words • 0.0% coverageAjay Malik • 14 words • 0.0% coverageAjay Malik • 36 words • 0.0% coverageAjay Malik • 22 words • 0.0% coverageAjay Malik • 10 words • 0.0% coverageAjay Malik • 19 words • 0.0% coverageAjay Malik • 4 words • 0.0% coverage
Selected voice

Thomas Erl

100%flagged-word coverage
66 attributed words13% of attributed speech99% writer coverage
0%10.0%20.0%Unattributed Quote+19.7 ptsWriter: 0.0%Thomas Erl: 19.7%19.7%Biased Writer Voice-18.3 ptsWriter: 18.3%Thomas Erl: 0.0%0.0%Indoctrination-11.3 ptsWriter: 11.3%Thomas Erl: 0.0%0.0%Quote-first Misdirection-2.1 ptsWriter: 2.1%Thomas Erl: 0.0%0.0%

Attribution is sentence-level. Pattern percentages are calculated only from words assigned to that voice.

Loading…
Loading…
Loading…
Loading…

Analysis

Hover over highlighted words in the article to view the associated bias or fallacy analysis.