AI Backlash Is Really a Business Reality Check
Artificial intelligence is moving quickly from technology headlines into everyday business decisions. You may be seeing tools that promise to write marketing copy, build software, analyse documents, answer customers and automate back-office work. At the same time, a growing technology backlash is questioning whether these systems are as reliable, useful and accountable as their promotion suggests.
That tension matters if you run a Malaysian SME. You do not need to reject AI, but you also should not treat every impressive demonstration as proof that a tool is ready for your business. The practical question is simpler: can you verify the result, protect your customer data and recover quickly when the system gets something wrong?
What Happened
In a recent Decoder mailbag episode, The Verge examined why the current technology backlash feels different, particularly around AI, surveillance and the growing influence of large technology companies. The discussion highlighted a key distinction: software is comparatively easy to test because code can be run through a compiler or checked in a working environment. Businesses can often see whether a program produces the intended result. Source: The Verge, “Why the current tech backlash feels different”
Outside software, verification becomes much harder. A generated piece of code may work in a test environment, while an AI-generated customer explanation, legal summary, hiring recommendation or business forecast may sound convincing without being correct. The episode argued that enthusiasm for AI can become too broad when success in software is treated as evidence that the technology performs equally well in every other field. Source: The Verge, “Why the current tech backlash feels different”
The broader backlash is not simply opposition to new tools. It reflects concerns about whether people can understand how automated decisions are made, whether personal data is being used responsibly and whether companies are replacing human judgement before the technology is ready. For business owners, this is less about following an online argument and more about managing operational risk.
Why This Matters for Malaysian SMEs
Many Malaysian SMEs operate with small teams, limited time and several people performing multiple roles. You may be considering AI for customer service on WhatsApp, quotation preparation, social media content, invoice processing, stock updates or staff scheduling. These are sensible areas to explore because they involve repetitive tasks. However, each use case has a different level of risk.
For example, an AI assistant can help draft a reply to a customer asking about delivery times. You can verify that answer against your actual operating hours and courier arrangements before sending it. The same approach is less suitable for automatically approving a refund, interpreting an employment issue or giving advice about tax and compliance. In those situations, an incorrect answer can create a dispute or regulatory problem.
Malaysian businesses also need to think carefully about sensitive information. Customer names, phone numbers, identification details, payment information, supplier records and employee documents should not be pasted casually into public AI services. Malaysia’s Personal Data Protection Department provides guidance relating to personal data protection and compliance obligations, including the Personal Data Protection Act 2010. Source: Personal Data Protection Department Malaysia
You can start by separating information into categories. Public information, such as your published business hours, is generally easier to use with an AI tool. Internal information, such as sales reports and supplier terms, requires stronger access controls. Sensitive personal information should only be processed through tools that your business has reviewed, configured and approved.
| AI use case | Recommended approach | Verification needed |
|---|---|---|
| Marketing drafts | Use AI for first drafts and variations | Check claims, language and brand tone |
| Customer replies | Use approved answers and escalation rules | Confirm prices, stock and delivery details |
| Invoice data entry | Extract information into a controlled workflow | Match totals, tax details and supplier records |
| Recruitment screening | Use only as administrative support | Keep human review and consistent criteria |
| Contracts or compliance | Use AI to identify questions, not make final decisions | Ask a qualified professional to review the result |
Build Around Verifiable Work
The most useful lesson from the discussion is the importance of verifiability. Before adopting an AI feature, ask how you will know whether it worked. If an automated system reads an invoice, can someone compare the extracted amount with the original document? If it creates a product description, can you check specifications against your catalogue? If it classifies customer enquiries, can you measure whether urgent cases are being escalated correctly?
Do not ask only, “Can this AI do the task?” Ask, “How will you detect when it has done the task incorrectly?”
This changes the way you choose automation projects. Begin with a process where the expected result is clear and the error can be spotted quickly. For instance, you could use automation to move approved order information from a form into a spreadsheet or accounting workflow. You can then compare the source and destination records during a trial period.
Set an owner for every automated workflow. That person does not need to be a technical specialist, but they should know what the process is supposed to do, what information it can access and when a human must intervene. Keep a simple log of errors, customer complaints, missing records and unusual outputs. This gives you evidence about whether the system is improving the process.
The Bigger Picture
The technology backlash may continue because AI is being discussed as if every business problem can be solved by adding a chatbot or connecting a model to company data. In reality, automation works best when the underlying process is already reasonably clear. If your stock records are inconsistent, your customer information is duplicated or your approval responsibilities are unclear, AI may make the confusion move faster rather than remove it.
For Malaysian SMEs, responsible adoption is not about waiting until every uncertainty disappears. It is about choosing narrow, measurable applications and keeping people accountable for important decisions. You can use AI to reduce repetitive work while preserving human review for matters involving trust, safety, personal data, contracts and customer relationships.
The strongest businesses will treat AI as an operational tool, not a substitute for management. Start with one workflow, document the current process, define the acceptable result and test the system with non-sensitive data. Review its performance after a fixed period, then expand only if the evidence supports it. This approach lets you benefit from new technology without allowing hype to dictate how your business operates.
AI can help you respond faster and organise work more effectively. But the advantage will come from your process discipline: knowing what can be automated, what must be checked and who remains responsible when the answer is wrong.
Ready to Streamline Your Operations?
Technology moves fast. Your operations should keep up. AutoRunBiz builds AI systems that run your daily workflows — from WhatsApp order capture to accounting. Book a free 15-min ops audit →